Pulse Coded
Modulation
Jyotsna Bapat
Baseband Communication
2
Pulse Code Modulation
• Pulse Code Modulation (PCM) is the name given to the class of
baseband signals obtained from the quantized Pulse Amplitude
Modulated (PAM) signal by encoding each quantized sample into a
digital word.
• The source information is sampled and quantized to one of the L
levels; and each quantized sample is digitally encoded into an l bit
codeword (L = 2l).
3
4
5
6
7
Example
• Consider an analog input signal to PCM whose bandwidth is limited to 4 kHz and varies in
amplitude from -3.8 V to +3.8 V, with an average power of 30mW. The required signal-to-noise
quantization error ratio is given to be 20dB. Assuming uniform quantization, determine the
number of bits required per sample.
8
Example
• Signal power, P= 30mW
• required signal-to-noise quantization error ratio, (SNR) Q= P /σ2 = 20 dB= 100
• σ2 = 0.3mW =Δ2/12
• Δ=0.06 = (3.8+3.8)/2R
• R=7 bits.
9
10
“PCM is not an efficient system because it generates too many bits and requires so much
bandwidth to transmit”
Taylor Series Expansion
• Prediction of the (k + 1 )th sample from the two previous samples
Linear Prediction
• This is the equation of an Nth order predictor.
• Larger N would result in better prediction.
Differential Pulse Code Modulation
(DPCM)
• DPCM exploits the fact that most signals seen in practice are highly
correlated, especially if they are oversampled.
• Instead of transmitting the actual signal value, the difference between
current sample and its predicted value is quantized and transmitted.
• Since the difference is expected to be lower in amplitude compared to
actual signal values, the number of bits required to quantize can be
reduced (reduced bandwidth)
14
Simplified DPCM
• Simple DPCM transmitter and receiver diagrams as shown below:
15
Quantization Noise
• Quantization error is not limited to one sample (as in PCM).
• Quantization error may accumulate, resulting in considerable noise.
• Consider a sequence x[k] being transmitted using DPCM. Assume the quantizer used to quantize
d[k] is an 8–level quantizer with quantization intervals [–4,–3), [–3,–2), [–2,–1), … , [3,4) and the
output quantization levels are the center points in each interval (–3.5, –2.5, –1.5, … , 3.5).
16
Example
17
Differential Pulse-Code Modulation (DPCM)
Performance of DPCM is typically improved by using a prediction filter,
which predicts the value of the sample based on previous samples.
Summary: DPCM
• DPCM exploits the use of lossy data compression to remove the
redundancy inherent in a message signal, such as voice or
video
• It reduces the bit rate of the transmitted data (and, therefore,
transmission bandwidth) without serious degradation in overall
system response.
• However, this advantages come at the cost of increased system
complexity
Delta Modulation (DM)
• Delta modulation (DM), addresses another practical limitation of PCM, the need for
simplicity of implementation
• DM satisfies this requirement by intentionally “oversampling” the message signal
(typically 4 times Nyquist rate) increases correlation between adjacent samples.
• It results smaller prediction error that can be encoded using 1 bit (L=2).
• DM is basically 1-bit DPCM
• In effect, increased transmission bandwidth is traded off for reduced system complexity.
DM Transmitter
• In DM, we use a first-order predictor, which, is just a time delay of
Proceeding iteratively, and assuming zero initial
condition
(b) Delta modulator receiver
(a) Delta modulator
transmitter
E
How to control slope-overload
distortion?
• The slope overload noise can be reduced by increasing E (the step size)
• This unfortunately increases the granular noise.
• There is an optimum value of E, which yields the best compromise giving the
minimum overall noise.
• This optimum value of E depends on the sampling frequency fs and the nature of
the signal.
• No slope-overload distortion will occur if the staircase approximation follows a
steep segment of the message signal m(t).
Slope Overload
• Slope overload occurs when quantized signal cannot follow m(t). During sampling
interval Ts, mq^(t) is capable of changing by E (height of the step. Overload will not
occur if,
• Consider a case of tone modulation, the condition for no overload is
25
Example
• A sine wave signal with a frequency of 1 kHz is to be coded using a linear DM with a sampling rate
of 24 kHz. Find the step size required to encode the signal amplitudes of 100mV to ensure that
there is no slope overloading.
• Speech signal is bandlimited to 3 kHz and sampled at the rate of 8 kHz. To achieve the same
quality of distortion PCM requires 8 bits/sample and DPCM requires 4 bits/sample. Determine the
bit rates required to transmit the PCM and DPCM encoded signals
• A DM system is designed to operate at 3 times the Nyquist rate for a signal. The quantizing step
size is 250mV. determine the maximum amplitude of a 1kHz input sinusoid for which the DM does
not exhibit slope overloading.
26
